#fedora-meeting: Infrastructure (2012-07-19)
Meeting started by nirik at 18:00:00 UTC. The full logs are available at
* goedemiddag! (nirik, 18:00:00)
* New folks introductions and Apprentice tasks. (nirik, 18:01:25)
* Applications status / discussion (nirik, 18:05:00)
* Bodhi 2.0 meeting right after this meeting for interested parties.
* Sysadmin status / discussion (nirik, 18:17:24)
* Upcoming Tasks/Items (nirik, 18:23:02)
* 2012-07-19 migration of last redhat.com lists (smooge) (nirik,
* 2012-07-19 migration of lists.fedorahosted.org (smooge) (nirik,
* 2012-07-30 to 2012-08-03 PHX2 trip for smooge (tenative) (nirik,
* 2012-08-01 nag fi-apprentices (nirik, 18:23:13)
* 2012-08-07 to 2012-08-21 F18 Alpha Freeze (nirik, 18:23:13)
* 2012-08-08 drop inactive apprentices. (nirik, 18:23:14)
* 2012-08-21 F18 Alpha release. (nirik, 18:23:15)
* 2012-08-31 end of 2nd quarter (nirik, 18:23:17)
* 2012-09-11 to 2012-09-25 F18 Beta Freeze (nirik, 18:23:19)
* 2012-09-25 F18 Beta release (nirik, 18:23:23)
* FAD news (nirik, 18:25:52)
* Open Floor (nirik, 18:26:32)
Meeting ended at 18:44:58 UTC.
Action Items, by person
People Present (lines said)
* nirik (82)
* skvidal (45)
* sgallagh (20)
* smooge (14)
* lmacken (10)
* GFLjay (9)
* relrod (8)
* pingou (7)
* threebean (6)
* zodbot (4)
* herlo (3)
* rossdylan (2)
* ingm4r (1)
* abadger1999 (0)
* ricky (0)
* mdomsch (0)
* dgilmore (0)
* CodeBlock (0)
18:00:00 <nirik> #startmeeting Infrastructure (2012-07-19)
18:00:00 <zodbot> Meeting started Thu Jul 19 18:00:00 2012 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:00 <nirik> #meetingname infrastructure
18:00:00 <zodbot> The meeting name has been set to 'infrastructure'
18:00:00 <nirik> #topic goedemiddag!
18:00:00 <nirik> #chair smooge skvidal CodeBlock ricky nirik abadger1999 lmacken dgilmore mdomsch threebean
18:00:00 <zodbot> Current chairs: CodeBlock abadger1999 dgilmore lmacken mdomsch nirik ricky skvidal smooge threebean
18:00:14 * skvidal is here
18:00:15 * ingm4r is here
18:00:15 * pingou
18:00:16 * lmacken
18:00:18 * rossdylan is here
18:00:22 <relrod> here
18:00:36 * threebean is here
18:01:21 <nirik> ok, lets go ahead and start.
18:01:25 <nirik> #topic New folks introductions and Apprentice tasks.
18:01:31 <nirik> If any new folks want to give a quick one line bio or any apprentices would like to ask general questions, they can do so now
18:01:51 <GFLjay> Hi there, I'm Jay Clark, from Raleigh-Durham, NC.
18:02:16 <nirik> GFLjay: welcome.
18:02:17 <GFLjay> I'm not new to linux, but I'm looking to turn my advocacy into activisim
18:02:44 <nirik> excellent. Are you more interested in sysadmin type tasks or application development? or both?
18:02:55 <GFLjay> both actually
18:03:07 <GFLjay> I'm a sysadmin by day, and moderate coder by night
18:03:24 <GFLjay> also going for a RHCE soon
18:03:40 <nirik> cool. We can get you setup in the sysadmin apprentice group and point you at some easyfix application tickets to look into after the meeting over in #fedora-admin.
18:03:59 <GFLjay> sounds good
18:04:14 <GFLjay> thanks for the welcome - I'll be here in the corner watching
18:04:16 <GFLjay> :-)
18:04:27 <nirik> excellent.
Welcome again... and do chime in with questions or comments anytime.
18:04:33 <GFLjay> k thx
18:04:34 <nirik> anyone else? or shall we move on?
18:05:00 <nirik> #topic Applications status / discussion
18:05:33 <nirik> abadger1999 / threebean / lmacken / pingou / ianweller / relrod (and anyone I forgot): any new application news this week or upcoming?
18:05:51 <pingou> not much for me
18:05:52 <lmacken> We're going to be having a Bodhi 2.0 planning meeting after this, so if you're interested, stick around.
18:06:00 <pingou> might be more next week
18:06:10 <rossdylan> fedbadges is almost ready for staging, I just need to rebuild the rpms
18:06:13 <nirik> #info Bodhi 2.0 meeting right after this meeting for interested parties.
18:06:15 <lmacken> Still blocking on legal for the openhw2012 app :
18:06:23 <nirik> lmacken:
18:06:26 <threebean> working on standing up busmon at https://apps.stg.fedoraproject.org/busmon -- working on websockets+haproxy now.
18:06:33 <relrod> not much from me, I've been doing sysadmin-ish stuff this week more than devel stuff.
18:06:38 * herlo is here
18:06:50 <nirik> ianweller: any news on the flask version of elections?
18:07:19 <skvidal> it's not webapp related but I've written some apps for our system provisioning - in this case builders
18:07:23 <lmacken> threebean: I can't wait to see busmon in action
18:07:25 * herlo notes that he'd like to get back onto stickynotes if there is still a need. I should have a day or two to dedicate to it next week.
18:07:28 <skvidal> http://fedorapeople.org/cgit/skvidal/public_git/scripts.git/tree/ansible/start-prov-boot.py
18:07:47 <skvidal> we're using that now for creating and reinstalling builders for koji (and eventually for not-koji)
18:07:50 <nirik> herlo: that would be great if you can work on it...
18:08:07 <herlo> nirik: k, I'll request a vm later on
18:08:16 <relrod> oh I guess there is that - Flask 0.9 is still in the process of being EPELified - one of the co maintainers just up'd the rawhide version of Flask to 0.9 recently. There's a new python-werkzeug in EPEL testing right now (first step toward flask 0.9) - if you have anyone has any werkzeug apps, please test them against epel testing's version
18:08:26 <nirik> herlo: ok.
18:08:41 <relrod> s/you have//g
18:08:49 <sgallagh> I've been starting work on putting together a Gerrit template in OpenShift that I'd like to make available to the FedoraHosted infrastructure.
18:08:59 <pingou> relrod: no flask itself in epel6?
18:09:05 <relrod> pingou: not yet, no
18:09:10 <nirik> also, somewhat application related, I have been working on cgit some. Now, pkgs, fedorapeople and fedorahosted all have it setup, and I have asked some folks to look into a nice css for us to theme it.
18:09:19 <relrod> pingou: it's being worked on.
18:09:28 <pingou> I got an app with unit-test so let me know when you have flask itself (and flask-wtf) and I'll run the tests
18:09:45 <nirik> sgallagh: cool. so projects can do their own gerrit's in openshift easily?
18:09:53 <sgallagh> nirik: That's the idea
18:09:56 <nirik> sgallagh: I wonder if we shouldn't add that kind of info to the hosted faq?
18:10:08 <sgallagh> Yeah, that's in the plan.
18:10:25 <lmacken> Oh, I enabled Python hash seed randomization in staging for all of our Python apps.
18:10:34 <smooge> cool
18:10:36 <lmacken> Please test them
18:10:37 <nirik> cool.
18:10:38 <sgallagh> It's not ready yet. I've been able to get Gerrit up and running on OpenShift, but I still need to write the template.
18:10:46 <nirik> lmacken: when do we want to push that to production? early next week perhaps? or sooner?
18:10:56 <lmacken> nirik: next week sounds good to me
18:11:08 <lmacken> there is only so much we can test in staging, sadly
18:11:19 <nirik> yeah
18:11:22 <lmacken> many code paths only run in prod, so we'll just have to enable it and see what explodes
18:11:35 <threebean> wild west
18:11:56 <nirik> If I someday get the cycles I would really like to revamp our releng/maintainer stuff in stg so it all actually works.
18:12:04 <smooge> sgallagh, what exactly is gerrit? I have heard it in the wikipedia list
18:12:35 <relrod> smooge: code review
18:12:36 <sgallagh> smooge: It's a powerful code-review tool built around (and atop) Git
18:12:38 <nirik> ie, from source to build in stg builder, to test updates pushes, etc.
18:12:47 <lmacken> nirik: yeah, that would be nice
18:12:49 <smooge> thnkx
18:13:01 <nirik> lmacken: yeah, just a lot of work, but should be possible someday.
18:13:30 <sgallagh> smooge: Reviews are submitted by 'git push' to the Gerrit server and can then be merged from the web interface into the master branch (if the Gerrit server is used as the master) or easily pulled into a committer's local checkout and pushed to the master from there.
18:14:10 <sgallagh> smooge: http://r.cyanogenmod.com is a good example of a well-trafficed Gerrit server
18:14:19 <nirik> smooge / ianweller: any thoughts on pushing mediawiki 119 to staging? want to let it get more epel testing time first?
18:15:07 <smooge> nirik, so far all the changes I need to make are for documentation. There is a report of a missing file that I need to check
18:15:19 <smooge> I should be ready by tomorrow
18:15:34 <threebean> heads up on that, I already have the fp-wiki in modules-staging for fedmsg testing. That being resolved back into production is blocking on setting up and testing the ssl-signing in stg.
18:16:13 <nirik> smooge: ok, would be cool if we could get it in production before freeze, but we can see.
18:16:14 <threebean> may or may not get in your way. I can probably just snapshot it and blow it away so you can move ahead.
18:16:49 <nirik> threebean: yeah, we can figure something out there. That was something thats not really solved too well in this model, but then the old one was just as bad at it.
18:17:10 <nirik> ok, shall we move on then?
18:17:18 * threebean nods
18:17:24 <nirik> #topic Sysadmin status / discussion
18:17:37 <nirik> so, new sysadmin news this week from anyone? or upcoming soon?
18:18:01 <nirik> I setup a new machine from a new sponsor, and relrod setup a proxy08/app08 on it.
18:18:22 <nirik> skvidal setup a billion builders.
18:18:27 <skvidal> so... the build spinning that I posted about to infra-list a while back
18:18:27 <relrod> yeah, those are just about finished being set up and will probably go live today unless anyone screams
18:18:34 <skvidal> has come to fruition
18:18:54 <skvidal> it takes between 8 and 10m to spin up, provision and get the builder in production
18:19:01 <nirik> we also did a mass reboot/update tuesday that didn't go so smoothly. ;(
18:19:01 <skvidal> and it's all handsoff
18:19:14 <skvidal> nirik: hopefully that's been resolved for the future
18:19:18 <nirik> yeah.
18:20:05 <nirik> smooge: any news on mailing lists stuff?
18:20:49 <nirik> and also smooge is heading out to our main datacenter at the end of the month.
18:20:50 <smooge> I have certs and am working on getting it working
18:21:24 <smooge> I can't get the sign-vault dealt with so I am going to focus on ml and mediawiki
18:21:38 <nirik> sounds good.
18:22:07 <nirik> also, as a FYI on our private cloud: we have hardware in... we just need to get it setup network wise and install it.
18:22:36 <nirik> not sure if we can get that done before smooge is out there or duing that time, but hopefully soon.
18:23:00 <nirik> so, speaking of upcoming...
18:23:02 <nirik> #topic Upcoming Tasks/Items
18:23:13 <nirik> #info 2012-07-19 migration of last redhat.com lists (smooge)
18:23:13 <nirik> #info 2012-07-19 migration of lists.fedorahosted.org (smooge)
18:23:13 <nirik> #info 2012-07-30 to 2012-08-03 PHX2 trip for smooge (tenative)
18:23:13 <nirik> #info 2012-08-01 nag fi-apprentices
18:23:13 <nirik> #info 2012-08-07 to 2012-08-21 F18 Alpha Freeze
18:23:14 <nirik> #info 2012-08-08 drop inactive apprentices.
18:23:15 <nirik> #info 2012-08-21 F18 Alpha release.
18:23:17 <nirik> #info 2012-08-31 end of 2nd quarter
18:23:19 <nirik> #info 2012-09-11 to 2012-09-25 F18 Beta Freeze
18:23:23 <nirik> #info 2012-09-25 F18 Beta release
18:23:26 <nirik> smooge: you want to push those lists things to next week? any particular day(s)?
18:23:38 <nirik> anyone else have items that they want to schedule or note ?
18:23:54 <smooge> I am going for Thursday next week.
18:24:04 <nirik> okey dokey
18:25:52 <nirik> #topic FAD news
18:26:03 <nirik> I've not gotten around to working on FAD planning much...
18:26:06 <nirik> hopefully soon.
18:26:32 <nirik> #topic Open Floor
18:26:38 <nirik> anyone have items for open floor?
18:27:25 <skvidal> nothing from me
18:27:31 <sgallagh> I'm considering sunsetting ReviewBoard as a hosted feature
18:27:38 <skvidal> sgallagh: who was using it?
18:28:04 <pingou> maybe we should, for the easyfix tickets maybe
18:28:05 <nirik> not too many projects...
18:28:21 <skvidal> does gerritt require local access to the git repos?
18:28:27 <skvidal> or can it operate out of a clone?
18:28:44 <pingou> skvidal: I think option 2) based on what has been said earlier
18:29:15 <sgallagh> skvidal: I think anaconda tried it for a while and gave up due to the performance issues.
18:29:27 <skvidal> so.. we'd need to run gerritt on fedorahosted itself?
18:29:40 <sgallagh> Other than that, I think only autoqa is using it
18:29:56 <sgallagh> skvidal: No, Gerrit could be hosted in OpenShift instead of in fhosted.
18:29:58 <skvidal> sgallagh: oh - you're talking about reviewboard
18:30:04 <sgallagh> (Added bonus, any openid will work)
18:30:13 <skvidal> sgallagh: okay
18:30:31 <sgallagh> skvidal: Sorry, I stepped away for a moment and missed some context
18:30:32 <skvidal> so this is a discussion we've had multiple times...
18:30:57 <sgallagh> skvidal: Well, I'm going to offer both Gerrit and ReviewBoard as OpenShift services instead of hosted in FI
18:30:58 <skvidal> but - do we want to restrict what and how we support fedorahosted?
18:31:15 <nirik> If we do the hosted lists migration next thursday, I think I might like to re-try the hosted03->01/01 migration again the following week
18:31:17 <skvidal> once we get the mailing lists off of hosted
18:31:19 <smooge> question.. how will people get to Gerritt via URL?
18:31:40 <skvidal> with mailing lists gone - we have 2 advantages, afaict
18:32:23 <skvidal> 1. project.fedorahosted.org can point to whatever system we want it to
18:32:23 <sgallagh> smooge: OpenShift has its own <project>.rhcloud.com URLs
18:32:27 <skvidal> 2. $project.fedorahosted.org does not need ALL of the fedorahosted data to be useful
18:32:57 <smooge> yes.. how does this affect someone going from *fedoraproject.org -> *rhcloud.org either SSL, or "message"
18:33:43 <skvidal> smooge: no ssl involved afaict
18:33:46 <skvidal> it's openid based
18:33:49 <nirik> skvidal: yeah, but we need some more work to make that happen... but yes, should be more possible
18:34:04 <sgallagh> smooge: If we do anything at all, I assume it would just be a redirect to the rhcloud site
18:34:24 <smooge> skvidal, ok I was wondering since I saw discussion earlier about some fedora app on rhcloud moving to fedora
18:34:25 <sgallagh> ID is handled by OpenID, so we don't need to tie it into FAS at all
18:34:40 <skvidal> smooge: that's openhw2012 - and that is about keeping the url fedoraproject.org
18:34:51 <skvidal> smooge: there doesn't appear to be any requirement for that for gerritt
18:35:16 <sgallagh> (Side-benefit: we don't have to force people to get a FAS account to contribute, like we did with fedorahosted.org/reviewboard)
18:35:19 <smooge> ok. I just wanted to make sure I understood why X does and Y doesn't.
18:35:31 <smooge> and that makes it clear
18:36:21 <skvidal> nirik: I think I'd love to get the set of steps necessary to shard out hosted
18:36:25 <skvidal> nirik: outlined a bit
18:36:32 <nirik> yeah, might be good...
18:36:34 <skvidal> nirik: since, at least to me, that makes sense to try and do
18:36:46 <skvidal> nirik: however, if you disagree, I'd like to know what our plans should be, etc
18:36:49 <nirik> I'd like to get it off drbd to gluster, so thats what I have been trying to get working
18:36:54 <smooge> hopefully we will have a wildcard for fedorahossted later this year
18:37:14 <skvidal> nirik: my reason for wanting to shard out fedorahosted is so we could, ideally, not HAVE to be inside serverbeach
18:37:22 <sgallagh> smooge: As I said above, I'd like to offer both Gerrit and ReviewBoard in OpenShift. I recently got python-django-socialregistration working with ReviewBoard, so it will be able to function with OpenID in OpenShift as well.
18:37:50 <sgallagh> That will be a major improvement, since something about the FAS tie-in was causing delays on the order of seconds.
18:37:52 <nirik> skvidal: right. I was hoping we could make a hosted-agilo01 and use that as a test bed perhaps?
18:38:03 <skvidal> sure
18:38:18 <nirik> then we could move those 2 projects that need it there and they can be our test subjects.
18:38:46 <nirik> once they are done we can migrate more as we like
18:39:10 <skvidal> nod
18:39:13 <skvidal> okay
18:39:26 <nirik> but I'd like to get off drbd soon... especially since it won't work with new kernels. ;(
18:39:37 <skvidal> nirik: so... right now
18:39:49 <nirik> I suppose we could just go to a single instance with local disk.
18:39:53 <skvidal> we're really only using it as a 'sorta backup'
18:39:56 <skvidal> yah
18:40:01 <skvidal> that's what I was wondering
18:40:07 <nirik> well, with gluster we can have both nodes serving.
18:40:10 <skvidal> would it make sense to just drop drbd and use the local disk?
18:40:19 <skvidal> I just meant
18:40:20 <nirik> so the load gets farmed out some too
18:40:26 <skvidal> until the gluster nodes are up
18:40:30 <skvidal> I don't mean forever
18:40:56 <nirik> as far as I can see the gluster pair is ready to go. No one has found anything in testing that has let me know anyhow.
18:41:14 <nirik> I just wanted to move lists off so we dont move them then move them again
18:41:17 <skvidal> nod
18:41:20 <nirik> before we move to the gluster pair
18:41:26 <skvidal> great
18:42:31 <nirik> so, I think 2012-07-30 in the evening might be ok... will see.
18:43:11 <nirik> ok, anything more for open flooring?
18:43:51 <skvidal> nothing leaps to mind.
18:44:56 <nirik> okey dokey. Thanks for coming everyone. Remember: there's a bodhi 2.0 meetup after this meeting in about 15min...
18:44:58 <nirik> #endmeeting
infrastructure mailing list